Workspace TINA

The Instant Notetaking Assistant

Notes that
keep up.

TINA listens to your microphone and the meeting itself, writes a live transcript, tells voices apart, and follows the conversation across languages.

  • 55 languages, one model. Switch mid-sentence without changing a setting.
  • Voices, not just words. Mic and meeting audio stay on separate channels for cleaner speaker labels.
  • Yours to keep. TXT and timestamped SRT, saved on your device — not on a server.

Start a session

Pick what TINA should listen to. You can change this next time.

Notes stay on this device. Only audio is sent to your transcriber.
01

Two ears

Keep your mic and the meeting’s own audio on separate channels. That means cleaner separation — even before diarization.

02

Speaker-aware

“You”, “Remote 1”, “Room 2”. Rename any voice once and every line updates across TXT, SRT, and notes.

03

Code-switches

English into Spanish into Japanese, mid-sentence. Hints are optional; the model automatically detects the language.

04

Exports that fit

Readable TXT for notes, timestamped SRT for video, Markdown for your wiki. All generated locally in one click.

Works where you do

A browser app, not another install.

TINA is a progressive web app. Install it from Chrome, Edge, Safari, or Firefox on desktop, Android, iPad, and iPhone. Desktop Chromium gives you tab and system audio; mobile browsers give you the microphone plus file import.

macOS Windows Linux Android iOS & iPadOS Chrome Edge Safari Firefox

Preferences

Settings

Transcription

Checking…
Checking setup…

Nova-3 multilingual with speaker diarization.

Stored only in this browser and sent directly to Deepgram. TINA never receives your key. A Member key is required for TINA Cloud tokens; a regular key works here.

Get a key

Appearance

Storage

—

Sessions and recordings live in this browser. Clearing site data or using a private window removes them.

About TINA

TINA is a local-first PWA. Audio is streamed to your configured transcriber; transcripts, notes, and optional recordings stay in this browser. No analytics, no accounts.

TINA 1.0 · Nova-3 multilingual ·

Take it with you

Export transcript

Preview —

      

Already recorded?

Import audio

The file is uploaded straight from this browser to Deepgram for transcription. TINA keeps only the transcript. Speaker separation is on by default.

Capture guide

Help & privacy

Capture meeting audio

PlatformMicrophoneMeeting / system audioBest path
Chrome / Edge · Windows, macOS, LinuxYesYes, tab or systemStart with both sources on
Safari · macOSYesUsually video onlyMic + import the recording
Firefox · desktopYesVaries by OSTry sharing, or import
Android · ChromeYesNot exposedMic + import the recording
iOS / iPadOS · SafariYesNot exposedMic + import the recording
  1. Start listening. Allow the microphone, then choose the meeting tab and turn on Share tab audio. Sharing an entire screen may offer Share system audio on Windows.
  2. Keep the meeting tab shared. If you stop sharing, TINA keeps recording your mic and fills the meeting channel with silence; you can stop and restart for a fresh connection.
  3. No tab audio in your browser? Import the recording your meeting app saved, or route system audio through a virtual device (BlackHole or Loopback on macOS, VB-CABLE on Windows) and select it as the microphone.

How speakers and languages are detected

TINA sends two channels when both sources are on. Channel 0 is your microphone, channel 1 is the meeting. Diarization then splits multiple voices inside each channel. In a recording with one mixed channel, diarization works across the whole file.

Nova-3 multilingual transcribes 55 languages and handles mid-sentence code-switching. A language hint biases detection but never restricts it.

Privacy, in plain words

  • Audio goes directly from this browser to Deepgram, the transcription provider. It is not sent to a TINA server.
  • Transcripts, notes, and optional recordings are stored in this browser’s IndexedDB. No account, no analytics, no cloud sync.
  • Your API key is stored in this browser’s local storage when you bring your own. On a TINA Cloud site, the browser receives a short-lived token instead.
  • Deleting a session removes its transcript and audio from this device. Clearing site data removes everything.

Recording laws vary. Get consent from everyone on the call before transcribing.

Shortcuts

N New session K Search sessions E Export current transcript Esc Close dialog or panel Space Pause / resume while live (outside inputs)

Install TINA

In Chrome or Edge, use the install icon in the address bar or the Install TINA button in the sidebar. On iPhone or iPad, open the Share menu and choose Add to Home Screen. Installed mode keeps the app shell available offline; transcription still needs a network connection.

Speaker label

Name this voice

This updates every line from the same voice, plus TXT, SRT, Markdown, and previews.